Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 49

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7fc998d9e828>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 49

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7fc9a07f6e10>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.0.0
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    inputs_real = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels), name='input_real')
    inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')
    learning_rate = tf.placeholder(tf.float32, (None), name='learning_rate')
    

    return (inputs_real, inputs_z, learning_rate)


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param images: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    alpha = 0.2
    with tf.variable_scope('discriminator', reuse=reuse):
        # images = 28x28x3
        x1 = tf.layers.conv2d(images, 128, 5, strides=1, padding='same')
        relu1 = tf.maximum(alpha * x1, x1)
        # 28x28x128
        
        x2 = tf.layers.conv2d(relu1, 256, 5, strides=2, padding='same')
        bn2 = tf.layers.batch_normalization(x2, training=True)
        relu2 = tf.maximum(alpha * bn2, bn2)
        # 14x14x256
        
        x3 = tf.layers.conv2d(relu2, 512, 5, strides=2, padding='same')
        bn3 = tf.layers.batch_normalization(x3, training=True)
        relu3 = tf.maximum(alpha * bn3, bn3)
        # 7x7x512
        
        # Flatten it 25,008
        flat = tf.reshape(relu3, (-1, 7*7*512))
        logits = tf.layers.dense(flat, 1)
        out = tf.sigmoid(logits)

    return (out, logits)


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    alpha = 0.2
    with tf.variable_scope('generator', reuse=not is_train):
        # 25,008
        x1 = tf.layers.dense(z, 7*7*512)
        
        # Reshape it: 7x7x512
        x1 = tf.reshape(x1, (-1, 7, 7, 512))
        x1 = tf.layers.batch_normalization(x1, training=is_train)
        x1 = tf.maximum(alpha * x1, x1)
        # 7x7x512
        
        x2 = tf.layers.conv2d_transpose(x1, 256, 5, strides=2, padding='same')
        x2 = tf.layers.batch_normalization(x2, training=is_train)
        x2 = tf.maximum(alpha * x2, x2)
        # 14x14x256
        
        x3 = tf.layers.conv2d_transpose(x2, 128, 5, strides=2, padding='same')
        x3 = tf.layers.batch_normalization(x3, training=is_train)
        x3 = tf.maximum(alpha * x3, x3)
        # 28x28x128
        
        logits = tf.layers.conv2d_transpose(x3, out_channel_dim, 5, strides=1, padding='same')
        # 28x28x<out_channel_dim>
        out = tf.tanh(logits)
    
    return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    g_model = generator(input_z, out_channel_dim)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)

    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    all_update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)

    g_update_ops = [var for var in all_update_ops if var.name.startswith('generator')]
    d_update_ops = [var for var in all_update_ops if var.name.startswith('discriminator')]

    with tf.control_dependencies(d_update_ops):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1 = beta1).minimize(d_loss, var_list = d_vars)

    with tf.control_dependencies(g_update_ops):
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1 = beta1).minimize(g_loss, var_list = g_vars)
        
    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    input_real, input_z, lr = model_inputs(*data_shape[1:], z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, data_shape[-1])
    d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)
    
    samples, losses = [], []
    steps = 0
    sample_z = np.random.uniform(-1, 1, size=(72, z_dim))
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                batch_images = batch_images * 2
                steps += 1
                
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))
                
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z, lr: learning_rate})
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images, lr: learning_rate})
                
                if steps % 10 == 0:
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})
                    
                    print("Epoch {0}/{1}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:4f}...".format(train_loss_d),
                          "Generator Loss: {:4f}".format(train_loss_g))
                    losses.append((train_loss_d, train_loss_g))
                    
                if steps % 100 == 0:
                    show_generator_output(sess, 25, input_z, data_shape[-1], data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 50
z_dim = 100
learning_rate = 0.0005
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/2... Discriminator Loss: 0.000061... Generator Loss: 28.934143
Epoch 1/2... Discriminator Loss: 0.000021... Generator Loss: 27.642599
Epoch 1/2... Discriminator Loss: 0.010960... Generator Loss: 43.966309
Epoch 1/2... Discriminator Loss: 0.000649... Generator Loss: 19.246193
Epoch 1/2... Discriminator Loss: 0.026794... Generator Loss: 4.225798
Epoch 1/2... Discriminator Loss: 0.000145... Generator Loss: 12.476180
Epoch 1/2... Discriminator Loss: 1.562206... Generator Loss: 3.653619
Epoch 1/2... Discriminator Loss: 1.899490... Generator Loss: 0.381597
Epoch 1/2... Discriminator Loss: 0.000171... Generator Loss: 26.810593
Epoch 1/2... Discriminator Loss: 0.000656... Generator Loss: 18.872856

Epoch 1/2... Discriminator Loss: 0.330944... Generator Loss: 13.513503
Epoch 1/2... Discriminator Loss: 0.145003... Generator Loss: 4.630116
Epoch 1/2... Discriminator Loss: 0.209829... Generator Loss: 2.664790
Epoch 1/2... Discriminator Loss: 2.515746... Generator Loss: 0.218314
Epoch 1/2... Discriminator Loss: 1.148149... Generator Loss: 0.763774
Epoch 1/2... Discriminator Loss: 1.123564... Generator Loss: 1.149648
Epoch 1/2... Discriminator Loss: 3.200030... Generator Loss: 0.095254
Epoch 1/2... Discriminator Loss: 0.488049... Generator Loss: 1.555333
Epoch 1/2... Discriminator Loss: 1.177267... Generator Loss: 3.579447
Epoch 1/2... Discriminator Loss: 0.612334... Generator Loss: 2.504287

Epoch 1/2... Discriminator Loss: 1.289231... Generator Loss: 0.767128
Epoch 1/2... Discriminator Loss: 0.844882... Generator Loss: 1.622980
Epoch 1/2... Discriminator Loss: 1.029554... Generator Loss: 5.353817
Epoch 1/2... Discriminator Loss: 0.500920... Generator Loss: 2.226247
Epoch 1/2... Discriminator Loss: 1.587272... Generator Loss: 2.113319
Epoch 1/2... Discriminator Loss: 2.491234... Generator Loss: 0.234245
Epoch 1/2... Discriminator Loss: 0.618383... Generator Loss: 1.067074
Epoch 1/2... Discriminator Loss: 1.859493... Generator Loss: 3.263872
Epoch 1/2... Discriminator Loss: 1.844431... Generator Loss: 0.319788
Epoch 1/2... Discriminator Loss: 2.072435... Generator Loss: 2.847517

Epoch 1/2... Discriminator Loss: 1.959198... Generator Loss: 0.527076
Epoch 1/2... Discriminator Loss: 0.696382... Generator Loss: 1.744537
Epoch 1/2... Discriminator Loss: 1.094216... Generator Loss: 2.218685
Epoch 1/2... Discriminator Loss: 0.936725... Generator Loss: 2.121285
Epoch 1/2... Discriminator Loss: 1.171049... Generator Loss: 0.956687
Epoch 1/2... Discriminator Loss: 1.154859... Generator Loss: 0.698558
Epoch 1/2... Discriminator Loss: 1.516975... Generator Loss: 0.400964
Epoch 1/2... Discriminator Loss: 1.120958... Generator Loss: 0.757039
Epoch 1/2... Discriminator Loss: 1.399384... Generator Loss: 0.435594
Epoch 1/2... Discriminator Loss: 1.239990... Generator Loss: 2.399728

Epoch 1/2... Discriminator Loss: 1.136077... Generator Loss: 1.709656
Epoch 1/2... Discriminator Loss: 1.442653... Generator Loss: 1.982760
Epoch 1/2... Discriminator Loss: 1.192068... Generator Loss: 0.585033
Epoch 1/2... Discriminator Loss: 1.205437... Generator Loss: 0.544465
Epoch 1/2... Discriminator Loss: 1.080478... Generator Loss: 0.792274
Epoch 1/2... Discriminator Loss: 1.010549... Generator Loss: 0.699179
Epoch 1/2... Discriminator Loss: 1.449255... Generator Loss: 2.389877
Epoch 1/2... Discriminator Loss: 1.007452... Generator Loss: 0.759223
Epoch 1/2... Discriminator Loss: 1.168149... Generator Loss: 0.554722
Epoch 1/2... Discriminator Loss: 1.530747... Generator Loss: 2.346436

Epoch 1/2... Discriminator Loss: 0.820318... Generator Loss: 1.971665
Epoch 1/2... Discriminator Loss: 1.456802... Generator Loss: 0.347609
Epoch 1/2... Discriminator Loss: 1.288448... Generator Loss: 0.453849
Epoch 1/2... Discriminator Loss: 2.708148... Generator Loss: 3.609345
Epoch 1/2... Discriminator Loss: 1.192092... Generator Loss: 0.553120
Epoch 1/2... Discriminator Loss: 1.481587... Generator Loss: 0.385876
Epoch 1/2... Discriminator Loss: 1.631717... Generator Loss: 0.426208
Epoch 1/2... Discriminator Loss: 1.072385... Generator Loss: 1.256437
Epoch 1/2... Discriminator Loss: 1.423460... Generator Loss: 2.096963
Epoch 1/2... Discriminator Loss: 1.580127... Generator Loss: 2.151775

Epoch 1/2... Discriminator Loss: 1.328304... Generator Loss: 0.467984
Epoch 1/2... Discriminator Loss: 2.060570... Generator Loss: 0.213065
Epoch 1/2... Discriminator Loss: 1.286673... Generator Loss: 0.561041
Epoch 1/2... Discriminator Loss: 1.264662... Generator Loss: 1.997517
Epoch 1/2... Discriminator Loss: 1.025673... Generator Loss: 0.690524
Epoch 1/2... Discriminator Loss: 1.070693... Generator Loss: 0.991473
Epoch 1/2... Discriminator Loss: 1.211277... Generator Loss: 0.468890
Epoch 1/2... Discriminator Loss: 1.178902... Generator Loss: 1.622676
Epoch 1/2... Discriminator Loss: 0.919735... Generator Loss: 0.947386
Epoch 1/2... Discriminator Loss: 1.409069... Generator Loss: 1.980396

Epoch 1/2... Discriminator Loss: 1.764183... Generator Loss: 0.289077
Epoch 1/2... Discriminator Loss: 1.745019... Generator Loss: 0.292627
Epoch 1/2... Discriminator Loss: 3.047393... Generator Loss: 4.324504
Epoch 1/2... Discriminator Loss: 1.092608... Generator Loss: 1.355404
Epoch 1/2... Discriminator Loss: 1.538394... Generator Loss: 0.357498
Epoch 1/2... Discriminator Loss: 1.054472... Generator Loss: 0.712770
Epoch 1/2... Discriminator Loss: 1.882417... Generator Loss: 0.254200
Epoch 1/2... Discriminator Loss: 1.173761... Generator Loss: 0.652994
Epoch 1/2... Discriminator Loss: 1.570318... Generator Loss: 0.368578
Epoch 1/2... Discriminator Loss: 1.173203... Generator Loss: 1.435294

Epoch 1/2... Discriminator Loss: 0.981883... Generator Loss: 1.046257
Epoch 1/2... Discriminator Loss: 0.841305... Generator Loss: 1.435902
Epoch 1/2... Discriminator Loss: 2.073752... Generator Loss: 3.466253
Epoch 1/2... Discriminator Loss: 1.712069... Generator Loss: 0.371517
Epoch 1/2... Discriminator Loss: 1.178972... Generator Loss: 0.768791
Epoch 1/2... Discriminator Loss: 0.872356... Generator Loss: 1.845934
Epoch 1/2... Discriminator Loss: 1.392671... Generator Loss: 1.656019
Epoch 1/2... Discriminator Loss: 1.134521... Generator Loss: 0.599520
Epoch 1/2... Discriminator Loss: 1.018470... Generator Loss: 0.668882
Epoch 1/2... Discriminator Loss: 1.695413... Generator Loss: 1.908724

Epoch 1/2... Discriminator Loss: 1.860595... Generator Loss: 0.281783
Epoch 1/2... Discriminator Loss: 1.197059... Generator Loss: 0.600717
Epoch 1/2... Discriminator Loss: 1.029306... Generator Loss: 1.610611
Epoch 1/2... Discriminator Loss: 1.013761... Generator Loss: 1.448350
Epoch 1/2... Discriminator Loss: 1.513502... Generator Loss: 0.402169
Epoch 1/2... Discriminator Loss: 1.153992... Generator Loss: 0.603409
Epoch 1/2... Discriminator Loss: 1.000392... Generator Loss: 1.004187
Epoch 1/2... Discriminator Loss: 1.240092... Generator Loss: 0.823554
Epoch 1/2... Discriminator Loss: 0.735863... Generator Loss: 1.289168
Epoch 1/2... Discriminator Loss: 2.311208... Generator Loss: 0.162697

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 50
z_dim = 100
learning_rate = 0.0005
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 10

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/10... Discriminator Loss: 0.343487... Generator Loss: 37.018211
Epoch 1/10... Discriminator Loss: 0.035813... Generator Loss: 40.484802
Epoch 1/10... Discriminator Loss: 0.059457... Generator Loss: 42.867447
Epoch 1/10... Discriminator Loss: 0.538445... Generator Loss: 39.648499
Epoch 1/10... Discriminator Loss: 0.186596... Generator Loss: 8.103965
Epoch 1/10... Discriminator Loss: 0.100242... Generator Loss: 8.729134
Epoch 1/10... Discriminator Loss: 0.003299... Generator Loss: 8.169909
Epoch 1/10... Discriminator Loss: 1.052955... Generator Loss: 9.857722
Epoch 1/10... Discriminator Loss: 0.411124... Generator Loss: 5.235451
Epoch 1/10... Discriminator Loss: 0.127542... Generator Loss: 3.628533

Epoch 1/10... Discriminator Loss: 0.069815... Generator Loss: 10.734613
Epoch 1/10... Discriminator Loss: 0.024600... Generator Loss: 10.508707
Epoch 1/10... Discriminator Loss: 0.078405... Generator Loss: 4.545009
Epoch 1/10... Discriminator Loss: 0.990824... Generator Loss: 5.322680
Epoch 1/10... Discriminator Loss: 1.720766... Generator Loss: 0.778319
Epoch 1/10... Discriminator Loss: 0.849172... Generator Loss: 3.993165
Epoch 1/10... Discriminator Loss: 0.488929... Generator Loss: 2.107226
Epoch 1/10... Discriminator Loss: 1.663201... Generator Loss: 0.936162
Epoch 1/10... Discriminator Loss: 1.291642... Generator Loss: 0.804719
Epoch 1/10... Discriminator Loss: 1.289210... Generator Loss: 0.778875

Epoch 1/10... Discriminator Loss: 0.553217... Generator Loss: 2.087335
Epoch 1/10... Discriminator Loss: 0.830911... Generator Loss: 1.190981
Epoch 1/10... Discriminator Loss: 1.174801... Generator Loss: 1.853872
Epoch 1/10... Discriminator Loss: 1.669476... Generator Loss: 2.838577
Epoch 1/10... Discriminator Loss: 0.840488... Generator Loss: 2.005797
Epoch 1/10... Discriminator Loss: 0.513018... Generator Loss: 1.575963
Epoch 1/10... Discriminator Loss: 1.121901... Generator Loss: 1.162167
Epoch 1/10... Discriminator Loss: 0.615805... Generator Loss: 1.676747
Epoch 1/10... Discriminator Loss: 1.025697... Generator Loss: 1.649719
Epoch 1/10... Discriminator Loss: 0.469471... Generator Loss: 2.558455

Epoch 1/10... Discriminator Loss: 1.069663... Generator Loss: 1.108622
Epoch 1/10... Discriminator Loss: 2.767544... Generator Loss: 3.870247
Epoch 1/10... Discriminator Loss: 0.754061... Generator Loss: 2.222166
Epoch 1/10... Discriminator Loss: 6.324288... Generator Loss: 0.003877
Epoch 1/10... Discriminator Loss: 1.170883... Generator Loss: 0.621990
Epoch 1/10... Discriminator Loss: 1.220695... Generator Loss: 1.900558
Epoch 1/10... Discriminator Loss: 0.804288... Generator Loss: 1.096959
Epoch 1/10... Discriminator Loss: 0.743127... Generator Loss: 1.635123
Epoch 1/10... Discriminator Loss: 0.902319... Generator Loss: 0.969872
Epoch 1/10... Discriminator Loss: 0.829269... Generator Loss: 1.501406

Epoch 1/10... Discriminator Loss: 0.941336... Generator Loss: 0.825548
Epoch 1/10... Discriminator Loss: 0.937187... Generator Loss: 1.830943
Epoch 1/10... Discriminator Loss: 0.572275... Generator Loss: 1.643144
Epoch 1/10... Discriminator Loss: 1.359362... Generator Loss: 0.501783
Epoch 1/10... Discriminator Loss: 1.171597... Generator Loss: 2.623317
Epoch 1/10... Discriminator Loss: 2.866306... Generator Loss: 5.225190
Epoch 1/10... Discriminator Loss: 0.892604... Generator Loss: 0.894773
Epoch 1/10... Discriminator Loss: 2.911624... Generator Loss: 5.284602
Epoch 1/10... Discriminator Loss: 1.035323... Generator Loss: 0.780236
Epoch 1/10... Discriminator Loss: 0.986665... Generator Loss: 0.758678

Epoch 1/10... Discriminator Loss: 3.273005... Generator Loss: 4.827794
Epoch 1/10... Discriminator Loss: 0.836435... Generator Loss: 1.955009
Epoch 1/10... Discriminator Loss: 0.345340... Generator Loss: 2.711915
Epoch 1/10... Discriminator Loss: 0.728565... Generator Loss: 1.301618
Epoch 1/10... Discriminator Loss: 0.545251... Generator Loss: 2.909873
Epoch 1/10... Discriminator Loss: 1.496120... Generator Loss: 0.400748
Epoch 1/10... Discriminator Loss: 0.487695... Generator Loss: 4.283422
Epoch 1/10... Discriminator Loss: 0.963884... Generator Loss: 0.977784
Epoch 1/10... Discriminator Loss: 0.760781... Generator Loss: 1.009603
Epoch 1/10... Discriminator Loss: 1.340895... Generator Loss: 0.489475

Epoch 1/10... Discriminator Loss: 2.034837... Generator Loss: 5.070492
Epoch 1/10... Discriminator Loss: 1.021780... Generator Loss: 2.935746
Epoch 1/10... Discriminator Loss: 0.782128... Generator Loss: 1.328769
Epoch 1/10... Discriminator Loss: 0.234874... Generator Loss: 2.012713
Epoch 1/10... Discriminator Loss: 0.986381... Generator Loss: 3.251770
Epoch 1/10... Discriminator Loss: 1.023374... Generator Loss: 0.903618
Epoch 1/10... Discriminator Loss: 0.572901... Generator Loss: 1.381531
Epoch 1/10... Discriminator Loss: 1.174136... Generator Loss: 0.632241
Epoch 1/10... Discriminator Loss: 0.790021... Generator Loss: 3.181134
Epoch 1/10... Discriminator Loss: 0.539800... Generator Loss: 1.216355

Epoch 1/10... Discriminator Loss: 1.196892... Generator Loss: 0.773098
Epoch 1/10... Discriminator Loss: 0.380329... Generator Loss: 2.188738
Epoch 1/10... Discriminator Loss: 0.682279... Generator Loss: 1.645911
Epoch 1/10... Discriminator Loss: 0.345614... Generator Loss: 2.520242
Epoch 1/10... Discriminator Loss: 0.363088... Generator Loss: 2.680595
Epoch 1/10... Discriminator Loss: 1.531351... Generator Loss: 2.807475
Epoch 1/10... Discriminator Loss: 1.725686... Generator Loss: 0.285554
Epoch 1/10... Discriminator Loss: 0.969427... Generator Loss: 3.591451
Epoch 1/10... Discriminator Loss: 0.516948... Generator Loss: 1.845371
Epoch 1/10... Discriminator Loss: 0.809373... Generator Loss: 1.083680

Epoch 1/10... Discriminator Loss: 0.180876... Generator Loss: 3.992467
Epoch 1/10... Discriminator Loss: 1.257669... Generator Loss: 4.134929
Epoch 1/10... Discriminator Loss: 0.332083... Generator Loss: 4.901827
Epoch 1/10... Discriminator Loss: 0.965123... Generator Loss: 2.172289
Epoch 1/10... Discriminator Loss: 1.039028... Generator Loss: 1.476802
Epoch 1/10... Discriminator Loss: 1.401515... Generator Loss: 1.713700
Epoch 1/10... Discriminator Loss: 1.403426... Generator Loss: 0.682158
Epoch 1/10... Discriminator Loss: 0.648942... Generator Loss: 2.391026
Epoch 1/10... Discriminator Loss: 0.527674... Generator Loss: 1.644492
Epoch 1/10... Discriminator Loss: 0.883722... Generator Loss: 2.925871

Epoch 1/10... Discriminator Loss: 0.746537... Generator Loss: 2.945594
Epoch 1/10... Discriminator Loss: 0.761076... Generator Loss: 1.016970
Epoch 1/10... Discriminator Loss: 1.535024... Generator Loss: 4.628663
Epoch 1/10... Discriminator Loss: 1.718285... Generator Loss: 0.346569
Epoch 1/10... Discriminator Loss: 2.109766... Generator Loss: 2.750007
Epoch 1/10... Discriminator Loss: 0.942421... Generator Loss: 2.958986
Epoch 1/10... Discriminator Loss: 0.791004... Generator Loss: 0.989313
Epoch 1/10... Discriminator Loss: 0.511055... Generator Loss: 2.089428
Epoch 1/10... Discriminator Loss: 2.297536... Generator Loss: 4.182734
Epoch 1/10... Discriminator Loss: 0.811787... Generator Loss: 3.632585

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.